Ask A Cartographer

Accuracy of UTM Coordinate System

March 02 2010 | 2 comments
Categories: Map Data

I have been told that the UTM coordinate system is only accurate to 1 meter. Is this true? My understanding is that coordinate system accuracy is only expressed as a ratio such as 1 part in 10000.

Mapping Center Answer:

I think there might be two concepts being confused here. One is the accuracy of data collection or digitization. That would be expressed in terms of "accurate to 1 meter". This reflects the accuracy of the digital feature relative to its known location on the ground. For example, some recreation grade GPS units will specify that the accuracy of positioning is to within 10 meters, but some professional grade GPS units can actually collect points to sub-meter accuracy. Here is an ArcUser article that discusses this concept in more depth:

The other idea is the precision of the mathematics behind the map projections. For a map that uses the UTM coordinate system, the underlying map projection is a Transverse Mercator projection. U UTM zone has a scale factor of 0.9996 along its central meridian, or four parts in 10,000, which is equivalent to 1:2,500. That is to say, given the math behind this projection, if you measure 2,500 "somethings" (e.g., feet, meters, inches) within the UTM zone, it will not be more than 1 "something" (e.g., foot, meter, inch) from its true length. This is, as you note, a unitless ratio.

UTM inaccuracy could be a lot more than a meter posted by Michael Kennedy on Oct 12 2011 7:14PM
I calculate that, using the requirement of 1 part in 2500, a point in the southwest of a zone that touched the equator could have an error of 379.7 meters compared with a point in the northeast of that 8 degree tall subzone.
Confirmation posted by Melita Kennedy on Nov 22 2011 1:21PM
Michael, I'm sorry for the long delay in responding.

I still think you're mixing the positional accuracy with the distortion due to the map projection. Positional accuracy of a point could be anything from cm to a kilometer, but that's true in any map projection. Measuring a distance could certainly result in a length that is much more than a meter different from the length measured on the ground. The scale-at-a-point is measuring distortion due to the mathematics of the projection and is related to the projection plane which is based off the ellipsoid’s surface. If the area in question has significant elevation, distances and areas will be even more distorted from the ‘ground’ measurements. I had always taken the 1:2500 statement for a UTM zone as true. I was very surprised to see that a point at the edge of a zone and the equator is closer to 1 part in 1033. In fact, you have to be at 40N to get close to 1:2500 at a zone edge. Thank you for pointing this out!

If you would like to post a comment, please login.

Contact Us | Legal | Privacy |